Auspicious Birth Dates among Chinese in California Douglas Almond  Christine Pal Chee  Maria Micaela Sviatschi c  Nan Zhong Department of Economics  School of International and Public Aairs Columbia
238K - views

Auspicious Birth Dates among Chinese in California Douglas Almond Christine Pal Chee Maria Micaela Sviatschi c Nan Zhong Department of Economics School of International and Public Aairs Columbia

g the Beijing Olympics began at 808 pm on 882008 Given the potential for discretion in selecting particular dates of labor induction or scheduled Caesarean section Csection we consider whether ChineseAmerican births in California occur disproportiona

Download Pdf

Auspicious Birth Dates among Chinese in California Douglas Almond Christine Pal Chee Maria Micaela Sviatschi c Nan Zhong Department of Economics School of International and Public Aairs Columbia




Download Pdf - The PPT/PDF document "Auspicious Birth Dates among Chinese in ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.



Presentation on theme: "Auspicious Birth Dates among Chinese in California Douglas Almond Christine Pal Chee Maria Micaela Sviatschi c Nan Zhong Department of Economics School of International and Public Aairs Columbia"— Presentation transcript:


Page 1
Auspicious Birth Dates among Chinese in California Douglas Almond , Christine Pal Chee , Maria Micaela Sviatschi c, , Nan Zhong Department of Economics & School of International and Public Affairs, Columbia University, New York, NY, USA Department of Veterans Affairs, Menlo Park, CA, USA Department of Economics, Columbia University, New York, NY, USA School of International and Public Affairs, Columbia University, New York, NY, USA Abstract The number eight is considered lucky in Chinese culture, e.g. the Beijing Olympics began at 8:08 pm on 8/8/2008. Given

the potential for discretion in selecting particular dates of labor induction or scheduled Caesarean section (C-section), we consider whether Chinese-American births in California occur disproportionately on the 8th, 18th, or 28th day of the month. We find 2.3% “too many” Chinese births on these auspicious birth dates, whereas Whites show no corresponding increase. The increase in Chinese births is driven by higher parity C-sections: the number of repeat C-sections is 6% “too high” on auspicious birth dates. Sons born to Chinese parents account for the entire increase; daughter

deliveries do not seem to be timed to achieve “lucky” birth dates. We also find avoidance of repeat C-section deliveries on the 4th, 14th, and 24th of the month, considered unlucky in Chinese culture. Finally, we repli- cate earlier work finding that Friday the 13th delivery dates are avoided and document a particularly large decrease among Chinese. For Whites and Chinese in California, mothers with higher levels of education are particularly likely to avoid delivering on the 13th. Keywords: superstition, Chinese, eight, California, birth date. Corresponding author Email address:

mms2241@columbia.edu (Maria Micaela Sviatschi) Preprint submitted to Economics & Human Biology November 6, 2014
Page 2
1. Introduction Cultural preferences can exert a persistent effect on the fertility decisions of Asian immigrants to the West [1, 2, 3]. Following earlier work on sex selection in Asian countries, excess males births were found among Asian immigrants to Britain, particularly at higher parities[1]. In the US, excess male births among Asian sibships is accounted for by families where the first birth(s) are exclusively female[2, 3]. In the 2000 US Census 5%

sample, having a son is 50% more likely than the biological norm after two daughters when parents are Chinese, Korean, or South Asian race[2]. The authors interpret these patterns as driven by conscious decision making by parents[1, 2, 3]. 10 A potentially more benign cultural preference concerns auspicious dates of birth. For reasons expounded elsewhere[4], the number eight is considered lucky by many Chinese, and 4 unlucky. Birth dates falling on the 8th, 18th, or 28th day of the month can readily be achieved through a variety of means, including choosing the date of labor induction and

C-section (or postponement thereof). 15 Likewise scheduling of C-sections (or inductions) on the 4th, 14th, or 24th might be declined by parents in favor of adjacent dates. To our knowledge, it has not previously been considered whether births by Chinese are skewed to achieve an eight (or avoid a four). Previous work has considered whether births are timed vis a vis auspicious 20 birth years according to the Zodiac calendar [5, 6, 7, 8]. Conception timing and abortion play a large role in governing the effects for birth year, but cannot reli- ably achieve the “fine tuning” of birth

date considered in this paper. Thus, the mechanisms and consequences may differ. Additionally, whereas the supersti- tions regarding birth years are thought to be sex specific (“girls born in a specific 25 astrological year are regarded as less desirable” [8], usually the 1966 birth year), the Chinese eight and four superstitions per se should be gender neutral. Man- ifestation of these superstitions, however, may be gender-specific in the context of son preference among some Chinese Americans, which we consider below. By considering short -term changes in probability

of delivery method among 30
Page 3
Chinese, [9]’s analysis of births in Taiwan in 1998 is closest to our own. [9] found that the C-section rate was 14% higher on “auspicious dates”, where “auspicious” was not defined using 8s as here but rather “traditional cosmology and astrology” for determining dates “suitable for marriage”[9]. The extent to which the number of births were skewed to occur on such dates was not explicitly 35 considered. [10] found C-section deliveries are reduced in Taiwan during the “ghost month” of lunar July, when major surgical procedures may be

considered inauspicious. To our knowledge, it has not been considered whether Asian immigrants to the West show a preference for delivering on specific auspicious birth dates. 40 Previous research has found short-term manipulations to achieve desired dates of delivery among non-Chinese. Births drop 2-4% during obstetrics con- ferences [11], suggesting accommodation of physician schedules. Date discretion is also observed near the end of the calendar year, which confers a tax advantage for parents relative to birth in early January[12]. Likewise, births were delayed 45 in Australia to

receive a tax bonus[13]. Additionally, previous work has consid- ered whether certain dates considered unlucky in Western cultures are avoided. In Australia, there are 7.7% too few births on Friday the 13th[14]. In the US, the number of births fell 11% on Halloweens from 1996-2006 and increased 5% on Valentine’s Days[15]. 50 There is an extensive literature documenting variation in medical treatments that depart from clinical indication, including elective C-section[16] or cesarean delivery on maternal request CDMR. One motivation for CDMR is the “desire to plan/time delivery” [17]. Among

these non-clinical determinants of delivery method or delivery timing, achieving “auspicious” birth dates may be particu- 55 larly difficult to rationalize from the perspective of public health. That said, if there is an increase in the number of C-sections and births on auspicious dates, it is not clear whether it is the health care provider or the parents who drive such an increase [14]. To address this point, we will consider whether fetal gen- der affects the likelihood of having an auspicious birth date. Primarily through 60 prenatal diagnostic ultrasound, gender is routinely

revealed to parents prior to
Page 4
delivery in the US. Is achieving an auspicious date more likely when that child is male? Given previous findings of parental preferences for sons among Asian immigrants to the US, parents may be more keen to achieve auspicious dates for their sons. Unless healthcare providers likewise seek to deliver males (but not 65 females) on auspicious dates, it might suggest that it is parents (not providers) who are behind the skewed birth dates. 2. Methods 2.1. Study Design and Population We conducted a population-based cohort study using microdata

from indi- 70 vidual vital statistics natality records covering all live births in California from years 1991-2002, collected and maintained by the California Office of Statewide Health Planning and Development (OSHPD). The data we analyze are the same as in [18]. We chose California for the analysis because national natality data produced by National Center for Vital Statistics suppress exact date of birth 75 (beginning in 1989). Furthermore, California has a large population of Chinese Americans. OSHPD’a research database includes hospital discharge records linked to birth (and death)

certificate records. The birth certificate data report pregnancy and birth characteristics, including pregnancy and birth complications, birth 80 weight and gestational age, as well as parents’ age, educational attainment, and place of birth. While race of the newborn will be considered, Hispanic births were not separately identified after 1995. Approximately 96 percent of all births in the vital statistics records were successfully linked to discharge information, which includes admission, discharge date, as well as additional 85 treatment measures described in [18]

(measures not analyzed here). Hospital admissions up to one year after delivery are matched to the birth record for both mothers and infants. The initial dataset contains 6,762,921 births. We restrict the analysis to those births where the mother’s race is White or Chinese (82% and 2% of Cal- 90
Page 5
ifornia births, respectively). About 1% of birth records omit race, which we exclude. We also exclude records where gender is not reported. 2.2. Statistical Analysis We aggregate the microdata meeting above criteria by exact date of birth, gender, mom’s race (White or Chinese), mom’s

education, and procedure type 95 (primary C-section, repeat C-section, and induction). We conducted multivari- ate regression analysis using STATA (release 11) statistical software. In all the analysis, we include year ( ), month ( ), and day of the week ( DoW ) fixed ef- fects and main holidays effects (New Years Day, Memorial Day, Fourth of July, Labor Day, Thanksgiving, Christmas, plus Christmas Eve and New Years Eve) 100 in regressions of the form: Log#births y,m,DoW Contains8 y,m,DoW Contains4 y,m,DoW Holiday y,m,DoW Dow y,m,DoW (1) We are interested in estimates of and , the

coefficients on variables for whether day of month contains an 8 or 4 (respectively). We run separate regressions for Whites and Chinese. 3. Results 105 We begin by showing the number of C-sections per day by the day of the month (1-28) separately for Chinese and Whites in Figure 1 below (omitting dates greater than 28 because they do not occur in every month). On average, there are 2.3 births per day to Chinese parents in California and 110 per day for Whites. For Chinese, we see that the 8th, 18th, and 28th days of the month are 110 each associated with an increase in the number of

C-sections (red vertical lines, left panel). We see no corresponding pattern of increases on these 8 dates for Whites (red vertical lines, right panel). For both Whites and Chinese, we see a large drop in the number of births falling on the 13th of the month (green vertical
Page 6
lines), a pattern previously found elsewhere [14, 15, 19]. Our hypothesis does 115 not explain all of the day-to-day variation shown in the figures (e.g. the increase in the number of Chinese births on the 16th of the month is unexplained). To assess the statistical significance of these

patterns, we report regression results in Table 1 (while controlling for the other time effects described above). These regression results likewise indicate that Chinese births in California dis- 120 proportionately occur on the 8s and furthermore that this skewing is statistically significant. Column 1 shows a 2.3% excess in births falling on the 8s. Inter- estingly, this behavior is concentrated among male births. Column 2 shows a 4.1% excess in male births on dates with an 8. We also see a larger magnitude response for deliveries by C-section. There are approximately 7% “too

many 125 repeat male C-sections on the 8s (column 3). There is also some evidence of too few repeat C-sections on the 4s (roughly a 5% drop). Repeat C-sections for girls do not show corresponding pattern. Turning to the 13th, there is substantial evidence of avoidance among Chinese for male and female births alike. There are about 14% too few C-section births on 13th for boys and girls alike, and 130 suggestive evidence of an additional effect of the 13th when it falls on a Friday. Figure 1: Average number of repeated c-sections by day
Page 7
Table 1: Chinese Births (1) (2) (3)

(4) Contains8 0.0232** 0.0406*** 0.0691** 0.0160 (0.00977) (0.0140) (0.0289) (0.0292) Contains4 -0.0156 -0.00139 -0.0520* -0.0335 (0.0100) (0.0144) (0.0313) (0.0317) Thirteen -0.0623*** -0.0468* -0.137** -0.138** (0.0175) (0.0252) (0.0551) (0.0579) Friday13 -0.0373 -0.0992 -0.154 -0.245 (0.0468) (0.0673) (0.161) (0.175) Observa- tions 4,383 4,383 2,813 2,684 R-squared 0.255 0.147 0.075 0.061 Mom Chinese Chinese Chinese Chinese Birth All Boys Repeat CS-Boy Repeat CS-Girl We define the number of births in logarithmic scale. All models include year, month and day of the week fixed

effects as well as controls for holidays. Robust standard errors in parenthesis ***p 0.01, **p 0.05, *p 0.1 For Whites, there is no corresponding pattern with respect to the 8s and 4s. None of the point estimates reported in first row are distinguishable from 0 (and are under 1% in magnitude). There is, however, a consistent reduction in births on the 13th of the month. There also seems to be an even stronger 135 avoidance of Friday the 13th among deliveries that were repeat C-section. We find a 20-25% drop in C-sections for Friday the 13th (combing coefficients for the

13th and friday the 13th), an aversion which appears to be shared by males and females alike. In supplementary results shown in the appendix, we show that most of the 140 13 avoidance is concentrated among mothers with greater than a high school education. The coefficient on the interaction term of mother having more than a high school education and birth occurring on the 13th is consistently negative and large in magnitude (2nd row from bottom of appendix table). In contrast, 8-seeking (4-avoidance) does not seem to be concentrated among more or less 145
Page 8
Table 2: White

Births (1) (2) (3) (4) Contains8 0.00191 0.00208 -0.00769 0.00832 (0.00240) (0.00417) (0.00969) (0.00961) Contains4 0.00442* 0.00321 0.00304 0.0109 (0.00246) (0.00428) (0.00994) (0.00986) Thirteen -0.0223*** -0.0177** -0.101*** -0.110*** (0.00430) (0.00748) (0.0174) (0.0172) Friday13 -0.0224* -0.0246 -0.158*** -0.0881* (0.0115) (0.0200) (0.0464) (0.0461) Observa- tions 4,383 4,383 4,383 4,382 R-squared 0.915 0.781 0.855 0.870 Mom White White White White Birth All Boys Repeat CS-Boy Repeat CS-Girl All models include year, month and day of the week fixed effects as well as controls

for holidays. Robust standard errors in parenthesis ***p 0.01, **p 0.05, *p 0.1 educated Chinese. The California natality data also report the individual hour and minute of birth, which national data on births from NCHS data do not. In results not shown, we do not find any evidence of shifting of the time of day of the birth for either the hour or minute of birth (relative to Whites). Such precise birth 150 timing might be more difficult for parents to control than birth date. 4. Discussion Superstitious preferences have also been shown in other “life” decisions among Chinese,

including residential address choice in British Columbia.[4]. Shifting in birth timing might be worrisome in light of recent work finding neg- 155 ative health consequences for the newborn from accelerating deliveries, even short-term movements within “full-term” pregnancies[20, 21]. An irony of the birth date shifting being concentrated on Chinese-American sons is that Chinese-
Page 9
American daughters may be protected from its health consequences. That said, we do not detect systematic differences in birth weight associated with 8s or 4s. 160 The birth weight metric

has been criticized for not consistently capturing new- born health[22] and heterogeneous birth weight effects maybe be offset if some parents delay delivery and others accelerate it in response to lucky (unlucky) numbers. We did not detect significant changes in the rate of infant readmis- sion to hospital; because readmission is rare, our readmission estimates for lucky 165 (unlucky) dates among Chinese are imprecise. Whether these birth date shifts harm health or not, their observance suggests a previously undocumented di- mension of cultural persistence among some Chinese

Americans. That female Chinese newborns show no skewing towards eight in birth dates argues against delivery date effects being steered by health care providers. 170 Finally, the magnitude of the effects we estimate suggest that the vast ma- jority of Chinese in California do not time births to have auspicious birth dates. Likewise, the literature on sex preference among Asian immigrants finds most do not select sex. That said, Chinese sons delivered by repeat C-section in California are particularly likely to have their delivery date skewed by cultural 175 superstition. 4.1.

Limitations It is unclear to what extent switching into C-section delivery accounts for the effect we find. While we do observe whether delivery is vaginal and non-induced versus C-section, we cannot distinguish births that would have been C-section 180 on a nearby date (absent the four and eight superstitions). A large part of the date-shifting effect is presumably from C-sections whose dates were “merely relocated in time (and the delivery method unchanged). Thus, we do not think our application allows us to identify the causal effect of C-section on newborn health.

185 To our surprise, mothers with more education were more likely to avoid Friday the 13th delivery. Nevertheless, we do not know whether maternal edu- cation has a causal effect on adherence to this superstition, or instead something
Page 10
correlated with education (e.g. rapport with obstetrician) that accounts for the heterogeneity. 190 Acknowledgements The data used in this paper can be accessed by submitting an application to the California Office of Statewide Health Planning and Development. Please see [18] for additional dataset information. Financial support from

the National Science Foundation CAREER Award #SES-0847329 is gratefully acknowledged 195 (Almond). All opinions reflected in this article and any remaining errors are our own. 10
Page 11
Appendix Results by Mom’s Education (1) (2) (3) (4) Contains4 -0.0149 0.00407 -0.0156 0.00249 (0.0247) (0.00438) (0.0544) (0.0119) Contains8 0.0388 -0.000141 -0.0221 -0.00615 (0.0243) (0.00434) (0.0483) (0.0118) Thirteen -0.00689 -0.00781 0.102 -0.0397* (0.0435) (0.00777) (0.0941) (0.0211) Friday13 -0.154 -0.0253 -0.212 -0.120** (0.116) (0.0206) (0.138) (0.0560) MomHs 0.517*** -0.214***

0.0615*** -0.186*** (0.0115) (0.00207) (0.0216) (0.00562) Hs Contains4 -0.0152 0.00128 -0.0132 0.000630 (0.0342) (0.00614) (0.0680) (0.0167) Hs Contains8 -0.0410 -0.00307 0.0158 0.00183 (0.0341) (0.00614) (0.0621) (0.0167) Hs Thirteen -0.0689 -0.00856 -0.113 -0.0444 (0.0611) (0.0110) (0.118) (0.0299) Hs Friday13 0.304* -0.00167 0.0144 -0.0508 (0.161) (0.0289) (0.228) (0.0787) MomGths 1.921*** -0.0276*** 0.488*** -0.0685*** (0.0115) (0.00207) (0.0193) (0.00562) Gths Contains4 -0.00129 0.000199 -0.000806 0.0110 (0.0342) (0.00614) (0.0606) (0.0167) Gths Contains8 -0.0112 0.00855 0.0769 0.0173

(0.0341) (0.00614) (0.0545) (0.0167) Gths Thirteen -0.0718 -0.0361*** -0.330*** -0.159*** (0.0609) (0.0110) (0.107) (0.0299) Gths Friday13 0.107 0.0113 – 0.0418 (0.161) (0.0289) (0.0787) Observations 12,952 13,149 5,549 13,149 R-squared 0.756 0.822 0.253 0.815 Mom Chinese White Chinese White Birth All All Repeat CS Repeat CS All models include year, month and day of the week fixed effects as well as controls for holidays. MomHs indicates whether the mom has 12 years of education and MomGths whether she has more than 12. Robust standard errors in parenthesis ***p 0.01, **p 0.05, *p

0.1 The coefficient in column 3 is dropped due to the few number of observations for this category. 11
Page 12
References [1] S. Dubuc, D. Coleman, An increase in the sex ratio of births to india- 200 born mothers in england and wales: Evidence for sex-selective abortion, Population and Development Review 33 (2) (2007) 383–4000. [2] D. Almond, L. Edlund, Son biased sex ratios in the 2000 United States census, Proceedings of the National Academy of Sciences (PNAS) 105 (15) (2008) 5681–5682. 205 [3] J. Abrevaya, Are there missing girls in the united states? evidence from birth

data, American Economic Journal: Applied Economics 1 (2) (2009) 1–34. [4] N. M. Fortin, A. J. Hill, J. Huang, Superstition in the housing market, Economic Inquiry 52 (3) (2014) 974–993. doi:10.1111/ecin.12066 210 URL http://dx.doi.org/10.1111/ecin.12066 [5] K. Kaku, Y. S. Matsumoto, Influence of a folk superstition on fertility of japanese in california and hawaii, 1966, American Journal of Public Health 65 (2) (1975) 170–74. [6] D. Goodkind, Chinese lunar birth timing in singapore: New concerns for 215 child quality amidst multicultural modernity, Journal of Marriage and Fam- ily 58 (3)

(1996) pp. 784–795. URL http://www.jstor.org/stable/353736 [7] C. Rohlfs, A. Reed, H. Yamada, Causal effects of sex preference on sex- blind and sex-selective child avoidance and substitution across birth years: 220 Evidence from the Japanese year of the fire horse, Journal of Development Economics 92 (1) (2010) 82–95. URL http://ideas.repec.org/a/eee/deveco/v92y2010i1p82-95.html [8] Q.-T. Do, T. D. Phung, The importance of being wanted, Amer- ican Economic Journal: Applied Economics 2 (4) (2010) 236–53. 225 12
Page 13
doi:10.1257/app.2.4.236 URL

http://www.aeaweb.org/articles.php?doi=10.1257/app.2.4. 236 [9] J. C. Lo, Patients attitudes vs. physicians determination: implications for cesarean sections, Social Science & Medicine 57 (1) (2003) 91 – 96. 230 doi:http://dx.doi.org/10.1016/S0277-9536(02)00301-5 URL http://www.sciencedirect.com/science/article/pii/ S0277953602003015 [10] H.-C. Lin, S. Xirasagar, Y.-C. Tung, Impact of a cultural belief about ghost month on delivery mode in taiwan, Journal of Epidemiology and 235 Community Health 60 (6) (2006) 522–526. arXiv:http://jech.bmj.com/ content/60/6/522.full.pdf+html

doi:10.1136/jech.2005.041475 URL http://jech.bmj.com/content/60/6/522.abstract [11] J. S. Gans, A. Leigh, E. Varganova, Minding the shop: The case of obstetrics conferences, Social Science & Medicine 65 (7) (2007) 1458 240 1465. doi:http://dx.doi.org/10.1016/j.socscimed.2007.05.034 URL http://www.sciencedirect.com/science/article/pii/ S0277953607003176 [12] S. Dickert-Conlin, A. Chandra, Taxes and the timing of births, Journal of Political Economy 107 (1) (1999) 161–177. 245 [13] J. S. Gans, A. Leigh, Born on the first of july: An (un)natural experiment in birth timing, Journal of Public

Economics 93 (12) (2009) 246 – 263. doi:http://dx.doi.org/10.1016/j.jpubeco.2008.07.004 URL http://www.sciencedirect.com/science/article/pii/ S0047272708001217 250 [14] J. S. Gans, A. Leigh, Bargaining over labour: Do patients have any power?*, Economic Record 88 (281) (2012) 182–194. doi:10.1111/j. 1475-4932.2011.00776.x URL http://dx.doi.org/10.1111/j.1475-4932.2011.00776.x 13
Page 14
[15] B. R. Levy, P. H. Chung, M. D. Slade, Influence of valentine’s day and 255 halloween on birth timing, Social Science & Medicine 73 (8) (2011) 1246 1248.

doi:http://dx.doi.org/10.1016/j.socscimed.2011.07.008 URL http://www.sciencedirect.com/science/article/pii/ S0277953611004485 [16] H. Minkoff, F. A. Chervenak, Elective primary cesarean delivery, New Eng- 260 land Journal of Medicine 348 (10) (2003) 946–950, pMID: 12621140. arXiv: http://www.nejm.org/doi/pdf/10.1056/NEJMsb022734 doi:10.1056/ NEJMsb022734 URL http://www.nejm.org/doi/full/10.1056/NEJMsb022734 [17] E. J, Elective cesarean delivery on maternal request, JAMA 309 (18) 265 (2013) 1930–1936. arXiv:/data/Journals/JAMA/926917/jxr120011_ 1930_1936.pdf doi:10.1001/jama.2013.3982 URL

+http://dx.doi.org/10.1001/jama.2013.3982 [18] D. Almond, J. J. Doyle, After midnight: A regression discontinuity design in length of postpartum hospital stays, American Economic Journal: Eco- 270 nomic Policy 3 (3) (2011) 1–34. doi:10.1257/pol.3.3.1 URL http://www.aeaweb.org/articles.php?doi=10.1257/pol.3.3.1 [19] J. G. Miller, Revisiting the valentine’s day and halloween birth timing phe- nomenon, manuscript, Pomona College Department of Economics (2014). [20] A. T. Tita, M. B. Landon, C. Y. Spong, Y. Lai, K. J. Leveno, M. W. 275 Varner, A. H. Moawad, S. N. Caritis, P. J. Meis, R. J. Wapner,

Y. Sorokin, M. Miodovnik, M. Carpenter, A. M. Peaceman, M. J. O’Sullivan, B. M. Sibai, O. Langer, J. M. Thorp, S. M. Ramin, B. M. Mercer, Timing of elective repeat cesarean delivery at term and neonatal outcomes, New England Journal of Medicine 360 (2) (2009) 111–120, pMID: 19129525. 280 doi:10.1056/NEJMoa0803267 URL http://www.nejm.org/doi/full/10.1056/NEJMoa0803267 14
Page 15
[21] L. Schulkind, T. M. Shapiro, What a difference a day makes: Quantifying the effects of birth timing manipulation on infant health, Journal of Health Economics 33 (0) (2014) 139 – 158. 285

doi:http://dx.doi.org/10.1016/j.jhealeco.2013.11.003 URL http://www.sciencedirect.com/science/article/pii/ S0167629613001458 [22] D. Almond, K. Y. Chay, D. S. Lee, The costs of low birth weight, The Quarterly Journal of Economics 120 (3) (2005) 1031–1084. 290 15